致密气层压裂井产能规律研究

发布时间：2019-01-05 03:29

【摘要】：随着经济的迅速发展,天然气资源的需求越来越大,仅仅依靠常规天然气已经跟不上经济发展的脚步,越来越多的国家把目光投向致密气。但是致密气的特点不同于常规天然气,给开采和利用带来了很多难题。针对现场预测致密气层压裂井产能过程中出现的困难和不足,本文对有关致密气层压裂井产能的预测方法进行了相关研究。基于对致密气藏的地质特征和增产措施的研究,建立了致密气层压裂井产能影响因素的样本集。这主要是根据筛选原则对因素的影响程度进行定性和定量分析。采取数值模拟方式定性分析,减小了定量分析的复杂度。定量分析采取正交试验和灰色关联分析的方法,得到各个因素的具体影响程度。基于这个样本集,完善预测致密气层压裂井产能的模型。这里采取BP神经网络、支持向量机方法预测致密气层压裂井的产能,并且与解析公式结果对比。而投产后压裂气井的产量变化规律采取这两种方法和GM(1,1)模型进行预测。编制致密气层压裂井产能预测软件,并以某气藏的气井预测结果的准确性。通过对某气藏的气井进行分析发现,预测致密气层压裂井产能时,数学统计方法的准确率要好于解析公式法,其中支持向量机方法最高。预测投产后的压裂气井的产量变化规律时,支持向量机方法的准确率也是最高。支持向量机方法不需要大量样本数据,却有较好的预测效果。
[Abstract]:With the rapid development of economy, the demand for natural gas resources is increasing, only relying on conventional natural gas has been unable to keep up with the pace of economic development, more and more countries are looking to dense gas. However, the characteristics of dense gas are different from conventional natural gas, which brings many problems to exploitation and utilization. In view of the difficulties and shortcomings in predicting the productivity of tight gas reservoir fracturing wells in the field, this paper makes a related study on the prediction methods of the production capacity of tight gas reservoir fracturing wells. Based on the study of geological characteristics and stimulation measures of tight gas reservoirs, a sample set of factors affecting productivity of fracturing wells in tight gas reservoirs is established. This is mainly based on the screening principle of the impact of factors qualitative and quantitative analysis. The complexity of quantitative analysis is reduced by numerical simulation. The quantitative analysis adopts the method of orthogonal test and grey correlation analysis to obtain the specific influence degree of each factor. Based on this sample set, the model for predicting the productivity of tight gas reservoir fracturing wells is improved. In this paper, BP neural network and support vector machine are used to predict the productivity of tight gas reservoir fracturing wells, and the results are compared with the analytical formula. These two methods and GM (1 ~ 1) model are used to predict the production variation of fractured gas wells after putting into production. The productivity prediction software of tight gas reservoir fracturing well is compiled and the accuracy of gas well prediction result of a gas reservoir is given. Through the analysis of a gas well in a gas reservoir, it is found that the accuracy of mathematical statistical method is better than that of analytical formula method when predicting the productivity of fractured wells in tight gas reservoirs, among which the support vector machine method is the highest. The accuracy of SVM method is also the highest when predicting the rule of production variation of fracturing gas wells after putting into production. Support vector machine method does not need a lot of sample data, but it has good prediction effect.
【学位授予单位】：中国石油大学(华东)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE328

【参考文献】